In recent years, islet transplantation for diabetes has shown signs of the treatment efficacy, but its application is limited due to lack of donor organizations, sources and immune rejection. Bone marrow mesenchymal stem cells (BMSCs) have become a new resource of islet cell substitutes. One focus of the current research is the application of a specific inducing agent or a culture system to get directed differentiation of BMSCs, which may have part characteristics of islet cells and then be used in autologous transplantation for the treatment of diabetes.1,2 However, the results have been unsatisfactory. We selected nicotinamide and exendin-4 as induction agents,3–5 used different concentrations of glucose (high, medium and low levels) in culture media, and observed the difference in BMSCs induced effect. Meanwhile, cultured islet-like cells from the in vitro induction were transplanted into streptozotocin (STZ)-induced diabetic rats through different channels. Observation was made on their distribution in the rats and effects on blood glucose, so as to provide the theoretical basis for BMSCs induction into pancreatic islet-like cells and the function in the treatment of diabetes.
Induction of BMSCs
BMSCs of the Sprague-Dawley (SD) rats were separated and purified using the percoll density gradient centrifugation combined with the adherent culture method, followed by transmission and amplification. Under an inverted phase contrast microscope, they were observed on their changes in cell morphology. The third generation cells were harvested with a higher degree of purification for induction in Dulbecco's Modified Eagle Medium (DMEM). Experimental cells were assigned to the control group (without induction) and the induction groups (low glucose concentration (5.5 mmol/L, LG) group, medium glucose (11.1 mmol/L, MG) group, and high glucose (25.0 mmol/L, HG) group), which were, respectively, added with the inducing agents nicotinamide (10 mmol/L, Sigma, USA) and exendin-4 (10 nmol/L, Sigma) for 7 days and 14 days after being induced. Induction ended on the 21st day.
Identification of induced cells
Transmission electron microscopy (TEM) observation During differentiation, morphological changes of BMSCs were investigated under an inverted microscope. BMSCs and differentiated cells (D-MSCs) were fixed in 5% glu-taraldehyde for 2 hours at 4°C, transferred to 1% osmic acid for 2 hours at 4°C, and then dehydrated in acetonic acid. Ultra thin sections were counterstained using uranyl acetate and lead citrate, then viewed by TEM.
Dithizone (DTZ) staining
The cells were stained in freshly prepared DTZ solution and observed under a phase contrast microscope.
Cells were conventionally seeded. Rabbit anti-rat insulin antibody (Beijing Bioss Technology Development Co., Ltd., China) at the titers of 1:200 was taken as the primary antibody in accordance with the instructions of SABC kit (Wuhan Boster Biological Engineering Co., Ltd., China). DAB staining was followed. The standard for staining determination was based on brown-yellow granules as a positive outcome in cytoplasm. Image Pro Plus (IPP) was used for the image processing.
Cells were washed in PBS, and then add low-glucose-DMEM (5.5 mmol/L) for 12-hour stimulation, followed by collection of supernatant. High-glucose-DMEM (25.5 mmol/L) was added for the culture of 2 hours. Insulin radioimmunoassay kit (Beijing Puer Weive Bio Co., Ltd., China) was used for the determination of insulin content.
Gene ngn3 upstream primer (F): 5′-CGGGCAGAGCAG-ATAAAGC-3′, downstream primer (R): 3′-CGGGTGGT-CAGAGTCAAGC-5′, amplification length 215 bp; GK (F): 5′-GAGGGATCATCTGTGGGC-3′, (R): 3′-AGAAT TTTGTTTGCGGTCA-5′, amplification length of 172 bp; PDX-1 (F): 5′-ACCCGTACAGCCTACACTCG-3′, (R): 3′-GCTCGTTGTCCCGCTACT-5′, amplification length 201 bp; GLP-1 primer (F): 5′-CTGTCGGAGTGCG-AAGAGT-3′, (R): 3′-GGATGGCTGAAGCGATGA-5′, amplification length of 130 bp; GAPDH (F): 5′-TGCT-GAGTATGTCGTGGAG-3′, (R): 3′-GTCTTCTGAGTG-GCAGTGAT-5′, amplification length of 288 bp. PCR reaction conditions: 2 minutes at 50°C, 10 minutes at 95°C, 15 seconds at 95°C and 1 minute at 60°C, a total of 45 cycles.
Production of diabetes model
Twenty-four SD rats were injected with STZ (60 mg/kg) to destroy their pancreas and cause diabetes in rats. Three days later, the blood glucose was tested. Blood glucose ≥ 16.7 mmol/L and the stability lasted for 2 days. Meanwhile, rats were observed to have obvious symptoms such as polydipsia, polyuria.
The diabetic rats were assigned to: 1) renal cysts control group; 2) renal cyst group; 3) tail vein control group; 4) tail vein group. Islet-like cells (1×106) labeled with 10 μmol/L Brdu were transplanted from the rat tail vein or renal cysts into the bodies of diabetic rat models. Then the blood glucose of diabetic rats was observed in each group before transplantation and 3, 7, 14, 21, 28 days after transplantation. Twenty-eight days later, these rats were sacrificed. Immunohistochemical staining was used to determine the distribution of Brdu labeled cells and insulin in the heart, liver, lungs, spleen, stomach, kidneys, and pancreas. The control groups were injected with the cells without induction.
Data were expressed as mean±standard deviation (SD), and SPSS 11.0 (SPSS Inc., USA) was used for statistical analysis. Multi-factor analysis of variance for multiple groups followed by mean comparison with P ≤0.05 was s adopted to show the statistical significance.
After BMSCs were induced in LG, MG and HG medium, cells were in an elongated spindle shape under inverted phase contrast microscope, and showed the tendency to form cell clusters (Figure 1A). Cell clusters in HG-induced group were more than that in the others (Figure 1B). Meanwhile, compared with the control group (Figure 1C), the induced cells were presented with the typical features of secretory cells observed by TEM (Figure 1D).
The majority of cell clusters in induced groups were stained brown-red (Figure 2A), but not in the control group (Figure 2B). This showed that induced cell cytoplasm was rich in zinc ions.
The cells in the control group contained no brown-yellow cytoplasmic granules (Figure 3A); in LG group, some cells contained brown cytoplasmic granules (Figure 3B); in MG group, some cells had cytoplasmic brown granules more than that in LG group (Figure 3C); most of the cytoplasm in HG group had brown-yellow granules (Figure 3D). IPP image analysis showed that the value of integrated optical density in each group (A value) respectively: LG group (4.75±1.87), MG group (10.96±2.06), HG group (12.53±2.04), showing a significant difference between the groups (F=12.85, P <0.01). The value of integrated optical density (A value) of the control group (2.14±0.36) was significantly lower than that in the induced groups (P <0.01). This confirmed that induced cells were able to secrete insulin, and the HG group had higher number of positive cells than other groups.
Reaction of induced cells on glucose-stimulated insulin secretion
The low-glucose and high-glucose stimulated cells in the induced group were measured by radioimmunoassay and all found with insulin secretion-specifically, LG group: low-glucose stimulation ((2.73±0.20) IU/L), high-glucose stimulation ((3.95±0.24) IU/L); MG group: low-glucose stimulation ((3.43±0.38) IU/L), high-glucose stimulation ((4.90±0.38) IU/L); HG group: low-glucose stimulation ((4.57±0.22) IU/L), high-glucose stimulation ((6.64±0.46) IU/L), while the cells in the control group had no secretion. The differences between every two groups among them were significant (P <0.01). Under the high-glucose stimulation, the cells in HG group showed higher insulin secretion than that in the other groups. This confirmed that HG-induced group had a better effect.
Based on quantitative analysis of the expression of the pancreatic islet β-cell identity genes, islet β-cell-related genes GK, PDX-1, ngn3 mRNA were found to present significant changes on expression levels before and after induction (P <0.01). The level of each group increased as glucose concentration of the medium increased (P <0.01), while the ngn3 had the opposite expression (P <0.01); GLP-1 mRNA expression level varied in the control group and the induced groups (P <0.01), showing no trend associated with the glucose concentration.
Glucose and C-peptide before and after transplantation
Blood glucose and C-peptide of each group before and after transplantation are shown in Table.
Distribution of Brdu labeled cells and insulin-antibody
Detected by immunocytochemical staining, islet-like cells had Brdu labeled rate of 95% or more, and then immuno-histochemistry was employed to explore the distribution of Brdu labeled cells and insulin antibodies in various organs after transplantation. Renal cyst group could be observed with Brdu labeled cells in kidneys, with positive expression of insulin. In the tail vein group, Brdu labeled cells could be observed in pancreas, lungs, liver, kidneys and the corresponding expression of insulin was positive. The various organs in the control group were not found with that. Brdu positive color was mainly located in the nucleus, in brownish-black or brown. Insulin positive color was mainly located in the cytoplasm, in brownish yellow.
With the current upward trend in incidence, diabetes is a kind of disease that may seriously harm the human health worldwide. With the proposal of the Edmonton Program, islet transplantation has become one of the most valuable treatment means, but its applications are subject to the donor shortage and immune exclusion. Attributed to its self-renewal and multi-differentiation capacity, BMSCs are currently the most promising source of islet cells. For example, Jahr and Bretzel3 first used TCM-199 and nicotinamide to induce rat bone marrow mesenchymal stem cells into islet-like cells. Current researches are mostly focused on the choice of inducing agents at home and abroad.
This study was done from the micro-environmental perspective for the research on the glucose concentration of medium and its influence on the induction effect. Immunohistochemical staining showed that in HG group, most of the cytoplasm contained brown granules. Radioimmunoassay was used in measurement of insulin secretion. HG group had a higher insulin secretion than the other groups, confirming the better induction effect in HG group. It was suggested that after the induction, GK, PDX-1, ngn3 and the GLP-1 mRNA expression were significantly increased. It also found that, with the increase in glucose concentration in cultivation process, a synergy effect may occur to GK, PDX-1 mRNA expression, indicating apart from the role in inducing nicotinamide and exendin-4, medium glucose concentration can significantly influence the changes in gene levels of insulin-like β-cell. HG group had induced effects in the function and genetic level better than LG group and MG group. This proved that the increased medium glucose concentration can promote the production of induced β islet-like cells and the secretion of insulin, suggesting that in the process of in vitro induction into islet-like β-cell, HG environment is essential to the cells in vitro induced to be differentiated into islet-like cells.6–8 Glucose is not only a kind of cell nutrients, but also plays an important role in the induction for differentiation. Additionally, it functions as a major physiological regulator of insulin gene. Microenviron-ment may somewhat influence the orientation-induced differentiation of BMSCs. From the micro-environmental perspective, this provides a further theoretical basis for the BMSCs induced into islet-like cells.
The present study also found that islet-like cells induced by nicotinamide and exendin-4 may come through the tail vein and renal cysts channels, and then entered STZ rats to survive for a certain time, playing a hypoglycemic effect. Considering renal cysts as the immune escape zone,9 where cells can be protected from the immune attack, it is easy for the transplanted cells to feel the blood glucose changes and achieve the outward secretion of insulin. While in the tail vein group, it was found that the transplanted cells could be observed in lungs, liver, kidneys, pancreas and they had the expression of insulin, but no labeled cells were found in other organs. This is probably related to organ blood circulation, hemodynamics and histological features, and the specific mechanism remains unknown. On the 7th day in this study, the blood glucose began to decline, and dropped more on the 14th day. Until 28 days after transplantation, most of the blood glucose levels in diabetic rats had not been restored to the levels before modeling. C-peptide values after transplantation were significantly higher than that before transplantation. C-peptide and insulin were produced and released in an equal number of molecules in the islet cells, and not subject to the impact of exogenous insulin, thus being able to more accurately reflect islet β-cell functions. Renal cyst group and the tail vein group could both be detected with Brdu labeled cells and the insulin expression. This indicated that transplanted islet-like cells could secrete insulin in the body, resulting in a significant hypoglycemic effect, although the high blood glucose not completely reversed. These results are in line with the results of some other studies.10,11 In the study, we found that induced cells, after transplantation in rats, may reduce the STZ-induced high blood glucose, but induced cells can not be viewed as the islet β-cells in physiological sense in spite of islet cell-like characteristics. We still do not understand the mechanism of impact these cells may have. Future study should be done to find out how to improve the diabetic condition, and whether these cells can develop into mature β-cells.
In conclusion, BMSCS could be induced to differentiate into islet-like cells, similar to the insulin-secreting cells in morphology, structure, function and in protein and gene level. But whether it can play an in vivo role like insulin secretory cells requires further research.
1. Zhang W, Su BL, Meng XX, Liu B, Liu DD, Wang YY, et al. Induced in vitro
cell differentiation into insulin-secreting cells. J Clin Rehab Tissue Eng Res (Chin) 2008; 12: 3053-3056.
2. Yang NL, Yang F, Xu LL. Differentiation of adult bone marrow mesenchymal stem cells
into islet-like cells in vitro
. J Clin Rehab Tissue Eng Res (Chin) 2008; 12: 3053-3056.
3. Jahr H, Bretzel RG. Insulin-positive cells in vitro
generated from rat bone marrow stromal cells. Transplant Proc 2003; 35: 2140-2041.
4. Oh SH, Muzzonigro TM, Bae SH, Laplante JM, Hatch HM, Petersen BE. Adult bone marrow derived cells trans-diferentiating into insulin-producing cells for the treatment of type l diabetes. Lab Invest 2004; 84: 607-617.
5. Tang DQ, Cao LZ, Burkhardt BR, Xia CQ, Litherland SA, Atkinson MA, et al. In vivo
and in vitro
characterization of insulin-producing cells obtained from murine bone marrow. Diabetes 2004; 53: 1721-1732.
6. Kaczorowski DJ, Patterson ES, Jastromb WE, Shamblott MJ. Glucose-responsive insulin-producing cells from stem cells. Diabetes Metab Res Rev 2002; 18: 442.
7. Cao LZ, Tang DQ, Horb ME, Li SW, Yang LJ. High glucose is necessary for complete maturation of Pdx1-VP16-expressing hepatic cells into functional insulin-producing cells. Diabetes 2004; 53: 3168-3178.
8. Miyazaki S, Yamato E, Miyazaki J. Regulated expression of pdx-1 promotes in vitro
differentiation of insulin-producing cells from embryonic stem cells. Diabetes 2004; 53: 1030-1037.
9. Korsgren O, Jansson L, Andersson A, Sundler F. Reinnervation of transplanted pancreatic islets: acomparison between islets implan ted into the kidney, spleenor liver. Transplant Proc 1992; 24: 1025-1026.
10. Rajagopal J, Anderson WJ, Kume S, Martinez OI, Melton DA. Insulin staining of ES cell progeny from insulin uptake. Science 2003; 299: 363.
11. Blyszczuk P, Wobus AM. Stem cells and pancreatic differentiation in vitro
. J Biotechnol 2004; 113: 3-13.
Keywords:© 2010 Chinese Medical Association
rats; bone marrow mesenchymal stem cells; induction in vitro